We have used the Protein Interaction Statistics (PrInS) algorithm to statistically describe interactions between amino acids using protein structures. PrInS produces a scoring matrix to describe the frequency of amino acid interactions in the protein structures. The resulting scoring matrix was converted to an amino acids distance matrix M, where Mij value denotes the distance between the neighborhoods of amino acid i and j. Then, distance matrices were converted to rate matrices in order to be compared with the model substitution matrices, such BLOSUM62 and PAM120, in the prediction of protein evolution. Finally, outcomes were comparable, suggesting that our approach captures information about protein evolution process in a similar fashion as BLOSUM62 and PAM120.
Tetriary Structure of a Protein Molecule (Interactive Biology, 2017).